model farm program benefits: the citanduy
TRANSCRIPT
MODEL FARM PROGRAM BENEFITS: THE CITANDUY WATERSHED
By:
Bungsran Saragih Director, USESE
Paul C. Huszar Colorado State University
and
Harold C. Cochrane Colorado State University
July 1986
The opinions and views expressed in this paper are thoseof the authors and do not necessarily represent
the vieiws of USAID or any other agency.
EXECUTIVE SkMIMARI
The .Citanduy II project, which began in 1981/82 : intended to increase the incomes of farmers, reauce erosion, and increase employment opportunit±es in the uplana areas of the Citanduy River basin in kest and Central Java. The project established 48 model farm units and expansion areas (areas surrounaing the model farms) and introduced a variety of erosion control and other agricultural innovations including bench terraces, back-sloping, grassing terrace risers, channeling runoff, new cropping patterns and systems, and new input (e.g., seed, fertilizer,
insecticioe) mixes. Farmers were also given limited subsidies and credit to buy new inputs and constrwt their own terraces. To date, these improved agricultural practices have been introduced in over 10,000 hectares of critical land in the Citanduy upland areas.
USESbE, the Project Socio-Economic Evaluation bnit working with the Yayasan Agro-Ekonomi, and with assistance from raculty of the Department of Agricultural ana Resource Economics at Colorado State bniversity, conducted this study to determine the impact ot the improved agricultural practices on the project's beneficiaries. The study provides data on the performance of the model farm/expansion area component of the project and contributes to the rowing body of empirical data that can be used to improve our understanding of upland farming. The study's results will assist project decision makers in determining the appropriateness of the package of agricultural technology now being applied.
The study's main conclusions are:
I. Ihe actual value of production (crop yields x market price) averaged for the model farm and expansion areas increased from 136 ,000 rp/ha before terracing to 446,000 rp/ha after terracing. This increased
value of farm production is largely due to nhanging cropping patterns (for example, cassava production tell by 30%while rice and peanut
production rose by 23.7% and 14.4%, respectively).
i
2. Labor use (employment) w-th.n the model farm and expansion areas
Increased from an average ot 860 hr/ha before terracing to 1,774 hr/ha af :er terracing. The majority of increased labor use came from within the farmers' households.
3. Labor productivity (the value of production divided by labor use)
increased from an average of 158 rp/hr before terracing to 216 rp/hr
after terracinb.
4. The actual input mix of labor use and such purchased inputs as fertilizer, insecticide and seeds was suboptimal before terracing and rose to nearly optimal after terracing. This resulted in a
significant increase in the marginal value of both labor and
purchased inputs.
These findings demonstrate that with the adoption of appropriate
agriculture and conservation technologies, there is a significant potential for improving both farm incomes and rural employment
opportunities in the upland areas of Indonesia. However, before this program's practices and technologies can be applied on a wider scale,
three main issues need to be investigated further:
1. because the present study was conducted on sites with relatively
homogeneous agro-climatic data, little can be said about the relationship between agro-climatic conditions and production increases. A larger sar.ple with more diversity of agro-climatic
conditions is necessary in order to assess this relationship.
2. The present study does not include sufficient data to determine the full costs of the project. Understanding and measuring these costs is necessary to assess the need for and effectiveness of the Government of Indonesia offering incentives to terrace.
ii
3. A complete analysis requires an examination ol alternative methods of achieving the goals ot 'upland agricultural development. Thus, the costs and benefits ot agro-forestry and other methods of cropping on
aifterent slopes, soil types, and terracing methods, in adaition to bench terraces, must be investigated to aetermine the optimum program
mix.
it
Draft Translation
RINGKASAN PELAKSANAAN -RO CRAM
(EXECUTIVE SWbINARI)
Proyek Citanduy I, yang dimulai pada tahun 1981/1982, dimaksudkan untuk meningkatkan pendapatan para petani, mengurangi erosi tanah, dan meningkatkan kesempatan kerja di daerah dataran tinggi lembah sungai Citanduy di Jawa Barat dan Jawa Tengah. Proyek in( telah niembentuk 48 unit pertanian percontohan (motel farms) masing-masing dengan daerah perluasan (daeraih-aaerah disekitar unit pertanian percontohan tersebut) dan meniperkenalkan bermacam-macam cara pengendalian erosi tanah dan pembaharuan-pembaharuan dalam biang
pertanian termasuk terasering (bench terraces), pemiringan ke tengah (backsloping), perumputan tampingan teras (grassing terrace risers),
kanalisasi aliran air (channeling runoft), pola dan sistim penanaman baru, dan pola masukan baru seperti bibit, pupuk, dan pestisicla. Para
petani juga telah diberikan subsidi dan krecait dalam jumlah yang terbatas untuk membeli masukan .baru dan membangun teras milik mereka
sendiri. hingga saat in( teknologi baru pertanian tersebut telah diperkenalkan di atas tanah seluas 10 ,000 hektar yang keadaaunya
kritis di daerah dataran tinggi Citanduy.
USES1 (Unit Studi dan Evaluasi Sosial Ekonomi) telah bekerja sama dengan ayasan Agro-Ekonomi can cengan bantuan staf pengajar dari Department of Agricultural and Resource Economics di Universitas
Colorado untuk mengadakan penelitian guna menentukan dampak dari
telnologi pertanian yang baru pada penerlma manfaat proyek.
Penelitian in( memberikan data-data mengenai hasil-hasil car pertanian percontohan atau daerah perluasan dan ikut menunjang
berkembangnya data empiris yang nyata dan yang dapat dipergunakan
untuk meningkatkan pengetahuan kita mengenai pertanian dataran
iv
tinggi. hasil-fasil penelitian ini akan memantu para pengambil keputusan dalaw penentuan ketepatan dari paket teknologi pertanian yang kini sedang aiperhunakan.
hosimpulan utama dari penelitian ter&rbut adalah sebaga! berikut:
1. khilai konkrit dari produksi (hasil tani x hazga pasar) pertanian percontohan aan aaerah perluasan meningkat dari 136,000 rp/hektar sebelum aaanya terasering menjadi 446,000 rp/hektar. Ieningkatnya nilai hasil pertanian sebagian besar aisebabkan oleh adanya perubahrn pola penanaman (misalnya, basil ubi kayu turun sebanyak 30% sementara hasil beras aan kacang tanah masing-masing meningkat 23.7% dan 14.4%).
2. Penggunaan tenaga kerja dalam pertanian percontohan dan daerah perluasan meninkat dar-L rata-rata 860 jam/hektar sebelum terasering menjadi 1,774 jam/hektar. 1eningkatnya penggunaan tenaga kerja sebagian besar berasal aari keluarga petani.
5. Produktivitas tenaga kerja (nilai produksi dibagi pemakaian tenagakerja) meninfkat aari rata-rata 158 rp/am sebelum teraserin&menjadi
216 rp/jam.
4. Campuran masukan yang sebenarnya (actual input mix) antara pemakaian tenaga kerja dan masukan yang dibeli seperti pupuk, pestisida san bibit-Libit adalah dibawah optimal sebalum terasering cLan hampir mencapai titik optimal setelah terasering. Hal ini menimbulkan peningkatan yang berarti dalaw nilai marginal dari tenaga kerja dan masukan Yang dibeli.
Resimpulan-kesimpulan ini menunjukkan bahwa dengan pengunaan teknologi yang tepat untuk bidang pertanian dan pengawetan, ada potensi Yang tinggi untuk menintkatkan pendapatan pertanian dan kesempaten kerja di daerah dipeaesaan dataran tinggi di Indonesia. bagaimanapun juga, setelum pelaksanaan program dan teknologi ini dapat diteralkan lehth luas, tiga masalah pokok perlu ditinjau leblh lanjut, yaitu:
v
1. Oleh karena penelitian ini diadakan di daerah-daerah yang memnlAyai data iklim pertanian yang relatif homogen, tidak banyak yang bisa diutarakan mengenal hubungan danantara keadaan iklim pertanian peningkatan produksi. Contoh yang lebih luas dengan perbedaan iklim yang lebih menyolok diperlukan untuk menilai hubungan tersebut.
2. Penelitian yang sekarang ini tidak mencakup data yang cukup untuk menetalkan seluruh biaya proyek. biaya tersebut perlu diperhitungkan untuk dapat menilai kebutuhan aan keexektipan dari tawaran perangsang (untuk terasering) yang hendak aiberikan oleh pemerintah Indonesia.
3. bntuk analisa yang lebih lengkap aiperlukan acanya pengujian terhadap metode-metode pilihan demn tercapainya tujuan pen&embangan pertanian daerah dataran tinggi. Lengan demikian, biaya dan keuntungan-keuntungan ciar pertanian hutan (agro-forestry) dan metode-metode penanaman lain pada lereng dan tanah yang berbeda-beda jenisnya serta metode terasering harus diteliti terlebih aahulu untuk menetapkan pola campuran program yang optimal.
vi
CONTENTS
Page
EXEfLLIVE SUJiARY (ENGLISh) i
EXECUTIVE SU*1AR (BAHASA INDONESLA) iv
1. IRODULTION 1
1.1 Proj'ct History 1 1.2 Studies of the Citanduy II Project 2 1.3 Purpose of the Study 3
2. DESCRlPT1ON OF THE SINAGA STUIDY 5
2.1 Purpose 5 2.2 Overview of the Study Area 5 2.3 Methodology 6 2.4 1inaLings 7
3. FINDINGS 8
3.1 total Value of Production 10 3.2 labor Use 10 3.3 labor Productivity 11
4. ONLLJS1ONS 12
4.1 Causes o± Productivity Increases 12 4.2 The Contribution of lerracing to Increased
Production Values 12 4.3 Increases lar~inal Values of labor and
Purchasea Inputs 13 4.4 The Optimal ixes of Labor Use and Purchased Inputs
bse 13 4.5 Labor Productivity 14 4.6 Changin& Cropping Patterns 14
•5.REGCI0ILNDATIONS 17-
REFERENCES 19
APPENlIX A. ADD1IONAL SINAGA STbDD FINLINGS 20
APPENDIX b. kRODUU'lGN FUNCTION ANALISIS 23
APPENDIX C. REGRESSiON RESULTS: VALLE FUNCTIONS 32
PLBLICATiONS OF uENERAL INTEREST AVAILABLE IRON LSA]D/lNDGNESIA 34
vii
1. INTRODUCTION
1.1 Project History
The problems caused by the intensive cultivation of steep land
slopes in Java (erosion in the upper watersheds, flooding, and the
siltation of irrigation systems in the lower plains of most of the island's major river basins) have long been of national concern. lb
address these problems, the Government of Indonesia (GO1) launched the
Citanduy I project in 1976. Citanduy I was mainly directed toward
providing flood control, rehabilitating old irrigation systems, and
constructing new irrigation systems in the lowlands ot the Citanduy River
Basin in West Java.
In addition, the GO1 initiated a pilot project under Citanduy I
(called the Panawangan Pilot Project) to examine new solutions to the
root cause of lowland flooding and soil erosion resulting from the
intensive cultivation of steep slopes. In this project, the problems of
steep slope farming were approached from both an agricultural and a soil
conservation point of view, where efforts were made to develop farming
systems that increase production and income as well as incorporate soil
cons ervation techniques.
Based on the results of the Panawangan Pilot Project, the Citanduy
II project was begun in 1981/1982. The objectives of this project are to
increase the incomes of farmers, reduce soil erosion, and increase
employment opportunities in the upland areas of the Citanduy River Basin.
The core activities of Citanduy II are the establishment of 48 model
farm units and expansion areas over a five-year period.. The model farm
concept is an extension model which attempts to introduce a package of
upland agricultural technologies to farmer groups in contiguous areas of approximately 10 hectares. packageThe consists of bench terraces, new cropping patterns and systems, and new input mixes (e.g., seeds, chemical fertilizers, ana insecticides) for land with slopes of less than 45 percent. Land with slopes of more than 45 percent receive an
agro-forestry package. In addition to this relatively intensive set of extension activities, the project provides subsidies and credit to model-farm farmers over the tive-year period In order to allow them to construct bench terraces themselves, adopt new cropping patterns, and to buy new inputs.
Lxpansion areas are lands surrounding the model farms where land conditions are similar to those of the model farms. It was hoped that during the project, expansion area imitate thefarners would practices of the model farms. It this was done successfully, these extension area farmers were incorporated into the program and received input subsidies for three years and financial credit for their activities.
1.2 Studies of the Citanduy II Project
Several studies have been conducted to determine the success of the model farm program and the factors that contributed to this success. Sinaga et al. (1985) measured the project's impacts on three model farms that were built in 1981/82 and their extension areas. A similar study, with a larger coverage in terms of the number and age of model farms, was conducted by USESE (1986). Other studies of the Citanduy II model farm program include study by Irawana (1986), which examines the marketing of selected products of the model farms and their expansion areas, and a study by Tatuh (1986), which investigates credit for soil conservation
and farm practices.
2
All of the above studies investigate only the socio-economic aspects
o± the model farms and expansion areas. An important study of the agro-climatic conditions of land resources in the model farms area has been conducted by Xucera (1985). Rucera's study is intended to compliment the Sinaga and USESE studies.
1.3 Purpose of the Study
The original intention of the present study was to combine the Sinaga and ucera studies in a unified analysis of soil characteristics and to determine the project's beneficial impacts on the model farms and expansion areas. However, only a small subset of the data from the Kucera study car, be associated with farmers' data from the Sinaga study. In addition, the soil characteristics data that can be matched are relatively homogeneous. Therefore, only a limited analysis was possible.* bit the Sinaga data are rich in information and provide a basis for further investigation of the beneficial impacts of the model
farms program.
A preliminary analysis of the relation between production and
agro-climatic conditions found that the only significant relationships to the value of production are present land use and drainage. Land use has a positive relationship, but arainage appears to be negatively correlated. This negative relationship, however, may represent multicollinearity with some other unknown factor. The lack of a sufficiently large data base and the homogeneity of the data available are the likely causes of the failure to find a significant relationship between other agro-climatic variables and production.
3
This report contains the findings of a study conducted during a
one-month visit by Dr. bhngaran Saragih, Director of the Vait Studi and
Evaluasi Sosial Ikonomi (USESI), to the International School for
Agricultural and Resource Development (ISARD) of Colorado State
University. Given the data limitations noted above and the short period
provided to conduct the study, this analysis provides: (1) a verification
of the analysis conducted by Sinaga, (2) an investigation of the causal
relationships of the impacts observed by Sinaga, and (3) the
identification of future research needs. The next section of this paper
provides a brief description of the Sinaga analysis,
4
2. DESCRIPTION OF THE SINAGA STUDY
2.1 Purpose
The primary purpose of this study was to assess the socio-economic
impact of the Citanduy II project on model farm and extension area
tarmers. The focus of the study was limited to an examinatfon of la.d
and labor productivity and their related factors. Special attention was
also paid to the factors that may motivate iarmers to participate in the
project.
2.2 Overview of the Study Area
Site Selection
To ensure the most thorough assessment possible, three sites were
chosen using the criteria of selecting the project's oldeLt model farms
and selecting farms located in different subwatersheds (it was presumed
that sites located in different subwatersheds would also have different
agroclimatic conditions). The sites selected were: (1) Mekasari village
in the Cimuntur watershed area of Lest Java, (2) Cibahayu village in the
Citanduy Basin in West Java, and (3) Sadabumi village in the Clkawung
basin in Central Java.
Physical Ccnditions
The average elevations of the three sites range from 330 meters to 575 meters above sea level. Except in Cibahayu, where there are
considerable areas of flat land, the sites are characterized by moderate
to steep slopes. Although detailed information on soils was lacking, it
appears that these areas generally consist of volcanically-derived brown
latosols which are vulnerable to erosion. The sites' raintall ranges
from 2150 to 2320 mm per year.
5
Socio-economic Conditions
Statistical data on socio-economic conditions at these sites were
incomplete or lacking altogether. Available data indicated that
Cibahayu's population density is 1063 people per square kilometer and its agricultural density is 16.4 persons per square hectare. Comparable
figures for Sadabumi are 302 and 6.3, and for Mekasari, 510 and 5.2,
respectively.
The primary cash crops grown at the sites are coffee and clove trees
in Sadabumi, and coconut in ekasari, while much of the Cibahayu area is planted in bamboo. In addition, rice, maize, cassava, peanuts, various
annual vegetables, and bananas are the next-most prominent crops,
although manioc and squash are also grown. Land ownership is highly
fragmented, with farmers' plots located on as many as five different
blocks of land.
2.3 bmthodology
Selection of Sample Land Parcels and Farmers
Field blocks belonging to the project were randomly selected. Then,
because the owners ot each field block consisted of members of one farmer
group (Kelompok Tani), the sample plots were chosen through a random
selection of farmers, using membership lists of the selected field block farmer groups. Village membership lists were used to sample non-adopters
(see below).
Data Collection
Two basic techniques were used to collect data: the measurement of
fields or plots, and interviews with farmers. These two types of data collection were conducted simultaneously.
6
Number of Sample and Grcuping*
The population was divided into three groups: model farmers,
expansion area farmers (those who adopted the project's technologies witnin the expansion areas) and non-adopters. (because land ownership is fragmented into several plots in different blocks, non-adopters may have
participated in the program in blocks other than those for which they were surveyed. Thus, non-adopters here are farmers who elected not to
participate in the project in the blocks surveyed.) k total of 195 farmers comprised the sample. Ten model farmers in each of the three
villages, 40 extension area farmers in ead of the villages, and fifteen
non-adopters in each village were selected.
2.4 lindings
The main findings and conclusions of the Sinaga study are discussed
in Section 3. Additional details on Sinaga's findings are presented in
Appendix A.
* The original report used a slightly different terminology for these
farmer divisions, although their definitions have not chaiged here.
For clarity, the terms model farmers, expansion area farmers, and
non-adopters are u.ed consistently throughout this analysis.
7
3. FINDINGS
The results reported in the Sinaga (1985) study are based upon hand-calculations of preliminary data. Using micro-computers, we were
able to verify these results using a complete data set. In general, the
results of the Sinaga study apree with the computer-calculated values.
A table from the binaga study of particular interest to this study
is his Table 11; the recalculated version of this table is shown here as
lable 1. The original ana recalculated tables differ in two ways:
(1) Sinaga deflated his values by coverting rupiah values into rice
equivalents, but the recalculated values are left in nominal terms; and
(2) in addition to computing productivity values for the model farm and
expansion areas, we also computed these values for non-edopters.
8
Table 1. Land and Labor Productivity
Expansion Non-Adopter Model Farm Area Farm Farm
Ratio of Ratio of
After to After to Before After Before Before After Before Before
MEKASARI 1. Total Value per iectare 221,445 755,777 3.41 177,277 598,964 5.11 67,440
(rp/ha) 2. Labor Use 781 1,960 2.51 55,1 1,352 2.45 104
(hr/ha) 3. Labor Productivity 283 385 1.36 212 442 2.08 647
(rp!hr)
SADABUMI 1. Total Value per Hectare 166,349 398,385 2.39 99,222 251,195 2.53 206,705
(rp/ha) 2. Labor Use 743 2,911 3.92 902 1,496 1.66 963
(hr/ha) 3. Labor Productivity 223 136 0.61 109 167 1.53 214
(rp/ha)
CIBAHAYU 1. Total Value per Hectare 110,543 391,095 3.54 129,559 397,699 3.07 398,300
(rp/ha) 2. Labor Use 1,004 2,434 2.42 1,320 2,668 2.02 1,399
(hr/ha) 3. Labor Productivity 110 160 1.45 98 149 1.52 284
(rp/ha)
TOTAL SAMPLE 1. Total Value per Hectare 160,293 510,868 3.19 111,942 380,414 3.40 282,684
(rp/ha) 2. Labor Use 832 2,483 2.98 891 1,664 1.87 1,077
(hr/ha) 3. Labor Productivity 192 205 1.07 125 228 1.82 262
(rp/hr)
3.1 Total Value of Production
For both the model farms and expansion area farms in all villages, the project resulted in an increase in the total value of production (crop yields multiplied by market price): from an average of 136,000
rp/ha before terracing to 446,000 rp/ha after terracing. That iu, farm incomes increased by an average of 310,000 rp/ha/yr because of the
project.
On average, increase the ofthe in total value production was 219 percent in the model farm areas and 240 percent in the expansion areas.
However, non-adopters in Sadabumi and Cibahayu had greater returns per hectare than either the model farm areas or the expansion areas before terracing began (see Table 1), and non-adopters' total value of production in Cibahayu before terracing exceeded that of both model and expansion area farmers after terracing. Because of the non-adopters'
high initial value of production, their non-adoption of the model farm technology would thus to beappear economically rational. Nekasari's far higher adoption rate of project technologies appears to correspond to its relatively low value of production before terracing was introduced.
3.2 Iabor Use
Labor use (employment) within both the model farm and expansion areas increased from an average of 860 hr/ha before terracing to 1,774 hr/ha after terracing.
On average, the use of labor at model farms increased by 198 percent, and in the expansion areas it increased by 87 percent. After terracing, the model farm areas were found to use over twice as much labor per hectare and the expansion areas about one-and-one-half times as much labor as the non-adopters. The largest increase in labor use was a
I0
292 percent increase in the model farm area of Sadabumi and the smallest increase in labor use was the 66 percent increase iu the expansion area of Sadabumi.
3.3 labor Productivity
Labor productivity (the value of production divided by labor use) showed an overall increase, from an average of 158 rp/hr to 216 rp/hr before and after terracing, respectively.
labor productivity increased within both the model ferm and :e expansion areas of Mekasari and Cibahayu. It decreased in the model farm area and increased in the expansion area of Sadabumi. The largest changes in productivity occurred in 14ekasarl. On the average for all of the villages, labor Productivity in the model farms remained relatively constant at approximately 200 rp/hr, but increased by 82 percent in the expansion areas, from 125 rp/hr to 228 rp/hr. However, labor productivity for the non-adopters is, on the average, 262 rp/hr, thus providing another reason why these farmers are non-adopters.
11
4. CONCLUSIONS
The verification ot the Sinaga study (see Section 3) indicates that the total value of production (rp/ha), labor use (hr/ha) and labor productivity (rp/hr) all increased as a function of the model farm program. The purpose of this section is to examine the causes of these
increases. The results of this examination are presented here; Appendix B delineates the hypotheses and methodology used, and elaborates further
upon the results.
4.1 Causes of Productivity Increases
Changes in the input (labor use and purchased inputs) and output (crop) mixes account for the largest part of the increased value of productivity experienced in the project after terracing was introduced. Wihile terracing accounts for only a part of this increase, it must be given credit for facilitating some of the changes in input and output mixes. ast, the Citanduy II model farms program itself likely provided farmers with the information that was necessary to make these input and
output changes.
4.2 Thc Contribution of Terracing to Increased Production Values
-After the optimum mixes of labor and purchased inputs were computed (see Appendix B), they were used to calculate the maximum value of production per hectare. These inputs were then subtracted to comprise an on-farm income that combines both rents and on-farm wages.
Terracing was found to boost on-farm income for the average farmer by approximately 80,000 rp/ha/yr. thisbecause benefit does not include purchased input costs, which are subsidized under the program, the farmer
12
would receive an i.1ditional gain of 51,000 rp in subsidized purchased
input expenses under the project. This result was derived assuming optimal before- and after-terracin input mixes, so that the increased
returns are solely due to terracing.
Using a 12 percent discount rate over a 15-year period, the present
worth of these gains was calculated: 544,880 rp/ha without subsidies and 892,241 rp/ha with subsidies. Thus, to the extent that terracing costs
are less than 544,880 rp, it could be expected that farmers would voluntarily adopt terracing and that subsidies should not be required,
but credit may become very important..
4.3 Increased Marginal Values of Labor and Purchased Inputs
The marginal values of both labor and purchased inputs increased significantly with terracing. bor example, the marginal value of the 1,400th hour of labor increased from approximately 25 rp before terracing to 75 rp after terracing, assuming 10,000 rp of purchased inputs. Likewise, the marginal value of the 10,000th rp of purchased inputs
increased from approximately 0.7 rp before terracing to 4 rp after
terracing, assuming 1,400 hr of labor inputs.
4.4 The Optimal hixes of labor se and Purchased Inputs Use
After calculating the marginal costs of these inputs it was
determined that before terracing, 286 labor hours per hectare would produce an economically efficient solution for the farmer, and that 4,900
rp/ha was the efficient level ot application for purchased inputs. However, using a weighted average, it was found that model and expansion area farmers actually employed 860 hours of labor per hectare and that the value they received from the application of purchased inputs was only
25 percent of their costs.
13
The optimal after-terracing levels of these inputs were then calculated: 1,450 hours per hectare for labor use, and 51,100 rp of purchased inputs per hectare. After terracing, farmers were found to have achieved a near-optimal combination of labor and purchased inputs use: Their actual labor use was and1,774 hr/ha their actual purchased
inputs use was 57,400 rp/ha.
4.5 Labor Productivity
The productivity of labor increased from 1.43 hr/kg before terracing to 0.66 hr/kg after terracing for rice production, and from 2.5 hr/kg to 1.2 hr/kg for peanut production. For cassava production, labor
productivity remained relatively constant at 0.15 hr/kg.
4.6 Cbcnged Cropping Patterns
The price of each crop grown in the model and expansion farm areas was obtained by regressing physical and tarm revenue before and after the project's technologies were introduced. Applying these estimated prices
to the levels of production reported provided a breakdown of farm income by crop, and revealed that terracing has significantly altered the source
of incomes at these farms: Conservation measures like terracing have induced farners to shift from low-value crops such as cassava to crops
that earn three to five times more per kilogram (see Figure 1).
For example, cassava production fell fram 42.4 percent to 12.4 percent of the average value per plot, while rice production increased from 7.1 percent to 26.8 percent, and peanut production rose fram 3.4 percent to 17.8 percent. (Wne of the functions of terracing is to conserve moisture in the soil. The fact that terracing increased rice production by nearly 400 percent shows that this objective of the project
has been achieved.)
14
The most dramatic shifts occurred on slopes greater than 20 percent, where rice production increased trom 193 to 46,427 rp per plot, peanut production increased from 0 to 22,374 rp per plot, and cassava production
fell from 35,050 to 20,703 rp per plot.
15.
BEFORE TERRACING
CORN (4.3X)
OTHER (26.5X)
CASSAVA (42.4X
COFFEE (2.5X)
COCONUT (13.8X)
PEANUT (3.4X) RICE (7.1 X)
AVERAGE VALUE PER PLOT AFTER TERRACING
CASSAVA (12.4X) RICE (26.8X)
CORN (4.0X)
OTHER (10O.8%)
COFFEE (1.9X)
PEANUT (17.8%)
COCONUT (26.4%)
Figure 1. Average Value per Plot
16
5. REOMMENDATIONS
A number of important questions remain that cannot be answered by
this study. The questions and the necessary steps to answer them include:
1. Leta for the model farm program cover only three years. This is too short a period to project whether benefits will increase, decrease or
remain constant in the future. Therefore, further monitoring is
needed in order to assess the long-run potential of terracing.
2. The observed changes in cropping patterns contributed to the economic
viability of terracing. However, there may exist other cropping
patterns and perhaps even new crops that would result in even larger
gains from terracing. Research is needed to assess these
possibilities.
3. This study was only able to identify agro-climatic data with
relatively homogeneous characteristics for the model farms. As a
consequence, little can be said about the relationship between
agro-climatic conditions and productivity changes. A larger sample
with more diversity of agro-climatic conditions must be collected in
order to assess the importance of these conditions to the benefits
realized by the project..
4. The present study is confined to the on-farm benefits o± terracing,
but there may also be significant off-farm benefits downstream from
reduced soil erosion (including reduced sediment volume, improved
water quality, reduced flooding, etc.) Thus, both the quantification
of these off-farm benefits and an examination of the role of
extension activities in influencing farmers' decisions to adopt
conservation methods are necessary to evaluate the model farm program
fully and are critical to policy decisions regarding subsidies for
the adoption of the program.
17
5. On the cost side, no data are available to assess the construction costs of terracing. Such informationI is necessary in order to determine the feasible range for private terracing costs, to evaluate government subsidy programs to induce terracing, and to assess the economic viability of the project. Actual cost data need to be collected for the model farms and expansion areas.
6. Also on the cost side, there may be hidden costs of management associated with the operation and maintenance of the terraces and waterways. Understanding and measuring these costs is necessary for determining the feasibility of private terracing and for assessingthe need for and success of government incentives to terrace.
7. The present study is confined to an assessment of bench terraces, but a complete analysis requires an examination of alternative methods of achieving the goals of increased productivity and reduced erosion. Investigation of the benefits and costs of agro-forestry and of other methods of cropping in different slopes and soil types is needed in orer to determine the optimum program mix.
8. Ihis study found that the observed mix of inputs (i.e., labor and purchased inputs such as chemical fertilizers) was suboptimal priorto terracing, but that after terracing the actual mix approached the optimal mix. Moreover, the output (crop) mix of farmers changed with terracing. Oae possible explanation of these changes is that the extension activities of the model farms provided improved infornation to farmers on optimum input and output mixes.
However, the available data are not sufficient to test this hypothesis. further research is needed to assess the role of extension activities in the production decisions of farmers.
18
RIYLRN LES
Irawan, lambang. (1986). Pemasaran lntuk Pengembangan Lsahatani lahan
Kering di DAS Citanduy. Ciamis, Indonesia: USISE.
Jucera, barl P. et al. (1985). hicro-TMdel Iarm Assessment of Land
Resources. Ciamis, Indonesia: RMi Inc.
Sinaga, Rudolb S. et al. (1985). beneficiary Impact Study Citanuy
Watershed Area: A Case of Three Model Farm Sites. Survey Report No.
02/SR/85. hogor, Indonesia: Agro-Economic Survey Foundation.
Tatuh, Jen. (1986). Kredit Untuk Conservasi Tanah dan Pengembangan
Usahatani Lahan Kering di Bagian Hlulu DAS Citanduy. Ciamis, Indonesia:
US hS E.
USES Tim. (1986). Studi Impak Sosial Ekonomi Model Farn di Wilayah DAS
Citanduy. Ciamis, Indonesia: USISE.
19
APPENDIX A
ADDITIONAL SINAGA STUDY FiNDiNGS
A.1 Cropping Contributions to Total Return per Hectare
Before the model farm technology was introduced, cassava contributed between 54 62and percent of the total per-hectare yield in the three villages, because of cassava's negative effect on soil conservation, however, growing this crop was not recommended, and sometimes prohibited, in the project. Instead, .arwers were encouraged to cultivate rice, corn, peanuts, vegetables, grass for livestock, perennials such as fruit trees, and other trees which could be used for tirewoods and industrial
materials.
After the model farm technology was applied, the role of cassava decreased drastically and increased return per hectare and an improved quality of agricultural produce resulted. In the Mekasari model farms, rice became the dominant crop and coconut ranked second. In the expansion area, coconut remained dominant. In the Sadabumi model farms, rice, peanut and coffee production increased, while rice became dominant in the expansion area, with cassava second in rank. In the Cibahayu model farm and expansion areas, corn remained the most important crop.
A. 2 Input Use
Before the model farm technology was introduced, none of the farmers in the three villages used high-yield variety seeds or pesticides. After the project, however, these purchased inputs were used by the model and expansion area farmers in all three villages.
20
Fertilizer use also increased under the project. Except in Sadabumi, where animal fertilizer seldom used and chemicalwas fertilizers were not used before the project, farmers increased the amount of both types of
fertilizer used.
21
A. 3. Employment Opportunities
During the project, labor use in all categories of farms increased, although the rate of increase varied considerably. In general, the rate of increase was higher in the model farms than the expansion farms. The overall proportion of family and hired laborers to labor use increase in the model tarms was 68 percent ana 32 percent, respectively; in the expansion farms, it was 85 percent and 15 percent, respectively.
In Mekasari, however, the greatest increase in labor use was for hired laborers. This was partly due to the type of perennial crop grown (coconut), which requires skill in harvesting.
Overall, increased labor use benefitted both males and females, with female laborers experiencing a slightly larger increase.
A.4 Comparison of Model, Exp..nsion Area and Nlon-Adopters' Farms
Except in Cit.:hayu, the value of production for the model and expansion area farms was higher after terracing than for the non-adopter farms. Labor productivity, on the other hand, was lower than the non-adopters productivity for all villages. Thus, it can be concluded
that despite the increased value of production experienced under the project, the model farm program increased employment opportunities but with little increase in marginal productivity. One possible reason for
the lower value of production in Cibahayu is that non-adopters were already employing cropping patterns similar to those recommended by the
program before terracing. In addition, Cibahayu non-adopters were using chemical tertilizers before the project began.
22
APPENDIX B
IRODUCTION FUNCTION ANALYSIS
B .1 Framework of the tnalysis
The central problem defined here is one of optimizing the application
of terracing to the upland regions. This optimization, of course, will
depend upon the benefits and costs of terracing. The benefits are
realized both on-farm, in the form of increased productivity, and
oft-farm, in the torm of reduced downstream sedimentation from erosion.
because data have not yet been collected on the off-farm benefits, this
analysis was confined to the on-farm benefits and costs.
A hypothetical set of costs and benefit functions are shown in Figure
B-I. The total private costs (TIC) of terracing are likely to be a
function of topography, because the steeper the slope, the greater the
volume of soil that must be moved and thus the greater the loss of
cultivatable land. That is, costs are likely to increase at an
increasing rate with steepening slopes.*
The total private benefits (TPB) of terracing can be expressed in
terms of additional net farm revenues stemming from changes in the
productivity of land, labor and/or purchased inputs, and from changes in
cropping patterns. Net farm revenues are simply the revenues earned
minus the cost of hired labor and purchased inputs, net of any
subsidies. Unlike the cost curve, there is no a priori reason to expect
that the benefit curve will take any particular shape. The benefit curve
in Figure B-1 assumes diminishing marginal returns rrom terracing as the
slope gets very steep, but increasing marginal returns for more gradual
slopes.
* A simulation model of terracing costs demonstrates increasing and
diminishing marginal returns with increasing volumes of soil terraced. No actual cost data are available. The simulated cost function cannot be compared with the benefit function at this time.
23
tPc
~TPB
CA,
4-.
InI
0 A'A B C C 3ope .4J.
of er apegB.1. Be ef t an Cost4-BFi ur
opeC C 0 'AB
Benefits and Costs of Terracing'Figure B.I.
24
Figure B-I indicates that terracing is optimizea for slopes equal to
b. However, farmers cultivating land with slopes ranging from A to C would likely also terrace because the private behefits exceed the total
private costs (TPC).
Summing off-farm erosion control benefits with the on-farm benefits
of terracing yields the total social benefit (ISB) curve shown in Figure
b-I. When off-farm benefits are included, the socially optimum range of
slopes that should be terraced is expanded by AA' on lower slopes and CC'
on higher slopes. Farmers on these slopes, however, will not terrace on
their own because the private benefits less than theare private costs.
lerracing on these slopes will require subsidies in order to make the
private net returns from terracing positive.
b.2 Estimation Methodology
Private benefits
The methodology employea to estimate private benefits was shaped to a
large extent by the data available. it would have been ideal if
production functions have beencould estimated for the individual crops,
and a model constructed to optimize the mix or labor, chemical, fertilizers, and insecticides, seeas, and land before ann after
terracing. The data providen only the value of production by sample
plot, the labor hours, and the value o. purchased inputs. At best, only
before ann after revenue functions could be estimated from this information. Given these limitations, the optimum mix of purchased and
labor inputs were computed before ana after terracing.
Value Functions
A Ccbb-Douglass form of value function was utilized to estimate the
value functions, as shown in equation (l):
25
= Lal INa 2VALUE exp(L) (1)
=where: VALUE the value of prortuction in rp per ha; =C the constant from the regression equation; =L labor hours uividea by ha; =al the coefficient attached to labor; =IN the purchaseu inputs measured in rp per ha; and =a2 the coefficient attached to the purchased inputs.
Value, ana labor inputs aivide by Thisthe and were plot area. implies constant returns for allto scale three factors, a restriction tlh-!t was toleratea Lecause sample areas to lessthe plot proves be than a hectare in almost all instances. Generalizing from these data in oraer to extrapolate to larger farms was consicierea too risky. Several supporting studies were conducted to determine whether tile constant returns argument could be justifiea. The results ,:rongly suggest that such an assumption
is warranted.
Orairary least square regressions were performed on the log transformations of value, and the labor and purchased inputs for both the periods before and after the terraces were constructed. The resulting
two revenue functions were then utilizea to determine the optimal combination o± inputs and output (value of production) before and after implen~nting the technology. The difference 'in output Value, given an economically efficient mix of labor and purchasea inputs, should provide
a consistent measure of the effect of terracing on upland agricultural regions. This strategy is superior to simply comparing the before and after yields per hectare, because productivity might have changed fcr reasons unrelated to the terracing. Other reasons for increases in productivity might include in theincreases efficiency with which factor inputs are used and employing new cropping patterns. The likelihood of mistaking these and similar enhancements to production f.pr the effects of terracing is diminished when the most eilicient approaches, before and after terracing, are used as benchmarks for the purpose of comarison.
26
Efficient Levels of Inputs
he tunctions represented by equation (1) provided the basis for
determining how each additional labor hour per hectare or rupiah of
purchased input per LLctare influences the value or procuction. This is,
of course, simply the first derivative of equation (1) with respect to L
and IN, as shown in equations (2) and (3). The optimum cowLination of
these ingredients is one which equates these marginal values to the
resource costs.
-iMPL = al exp(L) L(al l) INa 2 (2)
where: VMEL is the additional value (in rp) of an additional hour
of labor.
-VMIN = a2 exp(C) Lal IN(a2 1) (3)
where: VMPIN is the additional value (in rp) of an additional unit
of purchased inputs (in rp).
The wages for female and male workers were obtained by regressing the
hired wage bill against the hired worker hours. The resulting wage rates
and the marginal resource cost of purchasea inputs of 1 rp were used to
solve equations (2) and (3) for the optimal mix of inputs. These
solutions were then used in conjunction with equation (1) to determine
the optimum revenue for the average farm.
One problem inherent in this approach 7s which marginal resource cost
to charge for family labor. Nen, women, and childiren all have different
chores which are likely to differ from those assignea to off-farm
laborers. The wage bills includec in the data provided information about
payments to off-farm help. It is questionable whether these rates should
be applied to family members. No doubt the shadow price of on-farm labor
should be tied to what one could earn oft the farm. but the analysis is
27
never that simple. in the absence of any better iniornation, we assumed
that this shadow price was equivalent to that which a hired female worker
would earn. it was thought that the wages or hired males would reflect
heavy work which might not conlorm to the activities of the farm family.
The private benefits received by both the farm families ana hired
labor were compuLea by uultiplying the optimal labor hours on- and
off-farm by the marginal wage rate. Rents to the land owner Were derived
by computing the value of production using equation (1) and subtracting
the wage bill and purchased inputs. W~e assumed that subsiaies were not
necessary and that farmers could pay for fertilizers, seeos, ana
insecticides. 1he benefits receivea by the land owners are, therefore,
understated by the amount of subsidy the government wishes to offer. in
structuring the problem iii this way, we were able to determine whether
such subsiaies could be phasea out, that is, whether terracing was
worthwhile even without government involvement.
26
B.3 Results
Optimal and Actual Input Mix for All Plots
Prior to terracing, tarners employed labor up to the poiat where the value of an additional hour was 37.5 rp. This implies significant over-employment of labor, because the costs o1 an additional hired female and male worker are estimated to be 79 and 206 rp per hour, respectively. Assuming that the weighted averaged marginal cost of labor prior to terracing is 100 rp/hr, then 286 labor hours per hectare would produce an economically efficient solution. This is admittedly a small number and we know that on average model and expansion area farmers actually employed 806 hours per hectare before terracing. The reason for this 'discrepancy between what is considered eco nomically optimal and actual practice may lie in the shadow price attached to family labor:
100 rp/hr is likely too high. It is thus probable that significant under employment or unemployment exists, there by decreasing opportunity costs.
it also appears that prior to terracing, the value received from the application of purchased inputs is only 25 percent oL their cost, with the subsidy. The efficient level of application is only 4,90(. p/ha. One can only speculate as to why this might be the case. The xiditional analyses reported below provide at least a partial explanatio.
After terracing, the situation appears to have improved significantly. Figure b-2 shows the extent to which terracing has shifted the iacter demand curves. Farners achieved close to an optimum input combination after terracing. According to the estimated value function, the value-iaximizing levels of labor use and purchased inputs use are 1,450 labor hours/ha and 51,000 rp/ha, respectively. The average faruer employed 1,774 hours and 57,400 rp/ha, r:espectively. This is somewhat surprising because chemical inputs w-.re subsidized. Under such conditions, one would have expected that these Inputs would niave been
over-utilized. Perhaps the government limited the amount of subsidies to a level which either by accident or design pro(luced an economically
efficient solution.
29
MARGINAL VALUE OF OTHER INPUTS ASSUMING 1400 UNITS OF LABOR
2.8
2.6
2.4
2.2-
U
0 ki
I.
2
1.8 -
1.6
~.1 <z
1.4
1.2
0.6
0.4
0-I
0
I
10
0
I I I
20 ,30 (Thou=nds)
OTHER INPLTS(VALUE iN P) BEFORE + AFTER
III
40 50
320-'
MARGINAL VALUE OF LABOR ASSUMING 10,000 P OF OTHER INPUTS
300
280
260
l b0
240
220
200
180
140
120
100
80
60
40
2o
0 0
D
2
(rhousirbm) LABOR HOURS
BETORE + AFER
4
Figure B.2. Marginal Values
30
Optimum Farm Income Before and After Terracing
Given the efticient mixes of labor and purchased inputs reported
above, it was possible to compute the "best practice" solutions before
and after terracing. The private benefits to the tarmer were computed as
follows. The optimum mixes were used to attain the maximum value of production per hectare. Hired labor and purchased input costs were then
subtracted. The remainder comprised on-farm income that combines both
rents and on-farm wages.
Terracing boosts on-farm income (both rents ana implicit wages) by
79,983 rp/ha/yr. This benefit does not include purchased input costs, which are subsidizea under the program. The gain an average tarmer
perceives would thus be approximately 80,000 rp plus 51,000 rp in
subsidized purchased input expenses. The present worth of these gains,
using a 12 percent discount rate for 15 years, is 544,880 rp/ha without subsidies and 892,241 rp/ha with subsidies. To the extent terracing
costs are less than 544,880 rp/ha, one would expect farmers to voluntarily adopt the practice and subsidies should not be requirea, but
credit may become very important.
Cther benefits
The magnitude of the potential gains attributable to terracing led us
to examine whether this practice was enhancing the productivity of farm inputs given the same cropping patterns or whether it facilitated a
change in the mix of outputs. by regressing physical production against
farm revenue, we were able to obtain the price of each crop, before and
after the project's technologies were introduced. Applying these
estimated prices to the levels of production reported provided a breakdown of farm income by crop. ligure I in Section 4 reveals that the
program has significantly altered the source ot incomes. It has induced
farmers to shift from low valued crops such as cassava to those which
earn three to five times more per kilogram, particularly rice, peanuts
and coconuts.
31
APPENDIX C
REGRESSION RESULTS: VALUE FUNCTIONS
Coefficients
Description Constant LN (L/HA) LN (IM/HA) R 2 Sample
1 All Before 10.359 0.207 0.035 0.07 148 0.096 0.022
2 All After 7.232 0.484 0.171 0.23 148 0.137 0.051
3 Before Low Slope (10) 8.603 0.454 0.007 0.18 40 0.182 0.031
4 After Low Slope (10) 8.775 0.147 0.264 0.17 36 0.183 0.194
5 Before Int. ( 10, 20) 9.528 0.286 0.033 0.11 90 0.113 0.027
6 After Int. ( 10, 20) 7.025 0.529 0.158 0.18 97 0.207 0.066
7 Before High Slope (20) 8.216 0.524 0.011 0.15 26 0.264 0.099
8 After High Slope (20) 5.666 0.666 0.215 0.48 32
0.213 0.081
9 Mekasari Before 8.611 0.524 0.074 0.29 53 0.129 0.056
32
Description
13 Cibahayu Before
14 Cibahayu After
15 After Adoption 81
16 After Adoption 82
17 1 Revised--New Areas
18 2 Revised--New Areas
19 Small Farms Before
20 Small Farms After
21 Large Farms Before
22 Large Farms After
Constant
10.105
5.477
7.336
5.998
9.005
7.156
10.427
8.928
8.817
3.909
LN (L/IA)
0.085 0.128
0.112 0.149
0.101 0.211
0.701 0.175
0.378 0.092
0.496 0.136
0.186 0.167
0.241 0.167
0.469 0.137
0.869 0.251
LN (IM/HA)
0.J-53 0.023
0.565 0.145
0.455 0.198
0.107 0.053
0.018 0.021
0.169 0.051
0.051 0.031
0.189 0.056
-0.051 0.046
0.202 0,107
R 2
0.56
0.26
0.13
0.32
-.15
0.23
0.08
0.23
0.19
0.34
33
PUBLICATIONS OF GENERAL INTEREST AVAILAbLE IROM LSAlD/INDONISIA
.*A Preliminary View of indonesia 's L~nployment Problem and Some Optionsfor Solving It, by Robert L. Rucker, Septewoer 1985.
2. Demographic Lackground and Births Averten: Indonesian lamilyPlanning, 1980-i984, by John A. Ross, Terry hull, LuLu D. Cost, andLiL. Piet, ctober 1985.
3. Public Expenciture Impact: Education and Health, Indonesian FamilyPlanning, by Dennis N.W. Chao, John A. Koss, ann David L. Piet,October 1985.
4. A Survey of Private Sector Training in Indonesia, by Grant Cox, November 1985.
5. An Epidemiological Approach to Health Planning and Problem Solving:A Case Study frrm Aceh Province, Idonesia, Dy bteven L. 5oiter, 7li Azir Hasibuan, and burhanuddin Yusui, lebruary 1986.
6. Developing Manpower for Indonesia's National Family PlanningProgram: bKKbN's Experience with Overseas Graduate TraininglN83-i985, by Santoso S. Hamijoyo, Thomas X. 1'Agnes, ana Slamet Sudarman, April 1986.
7. The High Performnce Sederhana Irrigation Systems Project, by David4.kobinson, ay 1986.
8. The Sederhana Assessment Study for West Java, West Sumatra, NorthSumatra and South Sulawesi Provinces, Inonesia, by P.1. EXbA International Co. Ltd., Nay 1986.
9. Model Farm Program benefits: The Citanduy Watershed, by bungaranSaragih, Paul Haszar,. and Harold G. Codirane, June 1986.
34